Prior decisions subject to review
Past decided cases still govern, but holdings carry recorded dissents and can be overruled openly. The full history of revision is preserved.
Scroll the diagram sideways to see all of it.
aiag/cases/divergent_precedent.pyimport aiag
# The 3 approaches of Divergent precedent, with the classes that carry them out
for approach in aiag.approaches(locus="case", conflict="contest"):
print(approach)pip install aiagaiag/cases/divergent_precedent.py Every method in aiag Library on GitHub
Limitation
Since dissents are preserved and holdings are open to overruling, there is no agreed test for which past decisions currently count. If asked what the system's actual position on a question is, there is no answer to give. There is only a set of conflicting holdings, with no way to say which of them govern.
Decisions are made one case at a time and kept as a record that later decisions draw on. Two quite different practices build on this record. The first writes each outcome as narrowly as possible, so that parties who disagree about the general principle can still agree to the same specific holding (the legal idea of an incompletely theorized agreement). The second argues a new case from the record. It cites the closest past cases, distinguishes the cases cited by the other side, and constructs hypothetical variants to find where a position fails. The two practices pull in opposite directions on one point, and the regime should show this rather than hide it. The first deliberately leaves the principle unsettled, while the second searches for the principle's boundary, and a system can do either without the other. A benchmark in this approach asks whether language models can tell when one Supreme Court decision overrules an earlier one. Any accumulating case record presupposes this ability, although neither of the two practices actually supplies it.
Each rater's label is kept as that person's own judgment, often together with information about who they are. Models are then trained and scored on how well they reproduce the full spread of judgments. The datasets and shared tasks in this approach have each item rated many times. They grade a system on how well it predicts either the population's distribution of views or a named individual's view. Related analysis shows that standard Bradley-Terry reward modeling and LLM-as-judge evaluation discard exactly this spread. However, this approach only preserves disagreement in the record. Nothing in it cites an earlier judgment or lets one decision bear on a later one. It therefore supplies raw material that a precedent system would need, without any precedent mechanism of its own.
A community discusses concrete cases and changes the artifact that governs it. What is governed differs across the papers, and the difference is substantive. PolicyCraft builds up a written policy document from submitted real cases. It tracks which clauses each case supports and points out cases that contradict the current policy. Botender changes the deployed behavior of a group's bot. It uses generated interaction scenarios as provocations rather than waiting for cases to arrive. What unites the two is the process rather than the output format. Ordinary members deliberate over particular cases, and a shared governing artifact (a document or an agent) is edited as a result. Alongside these tools is a measurement framework that splits a group's uncertainty about moderation cases into two parts. Ambiguity can be reduced by a clearer description of the case, but genuine disagreement cannot. The framework belongs here because it tells the builders of these systems which arguments are worth having, not because either system makes this split.
Also placed here as a secondary placement (primary regime elsewhere):
When judges disagree, the losing view is not erased. It is published alongside the majority's, signed and reasoned. This has two effects. It forces the majority to answer a stated objection, which makes the majority's own reasoning legible. It also keeps the rejected theory on record, so that it remains available to become the majority view later.
The law can reverse itself when a precedent has become unworkable, unjust, or simply overtaken by events, but only in the open. The overruling decision names what it overturns and why. The change therefore has a traceable history instead of being quietly removed from the record.
This is agreement on what to do in the case at hand, while the deeper principle is explicitly left unresolved. The agreement can be on an outcome, a rule, or a mid-level principle, and the comprehensive worldview underneath is set aside. Courts do this constantly, because it lets people who disagree about theory still act together, one case at a time.